|Home | About | Journals | Submit | Contact Us | Français|
There is increasing evidence that patients who have problems with binge eating (BE) or BE disorder (BED) are quite common among the severely obese, including bariatric surgery candidates. The literature suggests that in many cases such eating behaviours improve after bariatric surgery, although this is not uniformly true. The current paper reviews the data on the development of BE, BED and loss of control (LOC) eating after bariatric surgery and the impact of these problems on long-term weight outcome. A search was made of various databases regarding evidence of BE, BED and LOC eating post-operatively in bariatric surgery patients. The data extracted from the literature suggests that 15 research studies have now examined this question. Fourteen of the available 15 studies suggest that the development of problems with BE, BED or LOC eating post-bariatric surgery is associated with less weight loss and/or more weight regain post-bariatric surgery. These data suggests that it is important to identify individuals at high risk for these problems, to follow them post-operatively, and, if appropriate interventions can be developed if such behaviours occur in order to maximize weight loss outcomes.
Obesity, defined as body mass index (BMI) ≥30, is increasing in prevalence in the US and globally (Moyer, 2012). It is estimated that approximately 34.4% of the US population was obese in 2007–2009 (Shields, Carol, & Ogden, 2011). Current treatments include pharmacotherapy, dietary therapy, behaviour therapy and physical activity for the obese (Recommendation on Screening; Am Fam Physician, 2012). Bariatric surgery is indicated for the morbidly obese (BMI ≥ 40) or those who are obese with a BMI ≥ 35 with at least two known associated adverse medical sequelae of obesity (Shirmer & Schauer, 2010). Depending on the type of operative procedure, both volume restriction and/or nutrient malabsorption contribute to significant excess weight loss, often accompanied by resolution of or improvement in obesity-related comorbidities. However, there is a subset of patients that either do not lose a substantial amount of weight or regain the lost weight at long-term follow-up (Frank & Crookes, 2010). Attempts to determine preoperative characteristics of the patients who have a worse outcome have been made, yet few clear predictors of such outcomes have emerged and been replicated (van Hout, 2005; Livhits et al., 2012).
A review of the extant literature reveals a dearth of longitudinal data for patients who have undergone bariatric surgery, although some long-term studies are currently underway (Belle et al., 2007; Sjöström, 2008). Several non-surgical factors have long been reported as influencing outcomes of bariatric surgery (Hsu et al., 1998; Sarwer, Wadden, & Fabricatore, 2005; van Hout, Verschure, & van Heck, 2005), and some attention has been focused on the presence of maladaptive eating patterns such as binge eating (BE) (Saunders, 1999; Green, Dymek-Valentine, Pytluk, le Grange, & Alverdy, 2004; Kalarchian et al., 2002; Wood & Ogden, 2012), BE disorder (BED) and eating with a sense of loss of control (LOC) (Burgmer et al., 2005; Colles, Dixon, & O’Brien, 2008). These have been associated with increased caloric intake, poorer outcomes and ultimately, failure of treatment (Bocchieri, Meana, & Fisher, 2002).
One of the areas of interest in the eating disorders, obesity and particularly in the bariatric surgery literature is the concept of ‘loss of control’ eating (Bocchieri et al., 2002). LOC eating is required in association with the usual criteria for BE in those with bulimia nervosa and/or BE disorder, and those problems generally requires the ingestion of an objectively large amount of food for an eating episode to constitute an eating binge. However, after bariatric surgery, the ingestion of an objectively large amount is difficult or impossible, although literature suggests that eventually, some post-bariatric patients can ingest an objectively large amount of food. The term LOC eating is now used at times instead of BE because of this limitation. Also, some authors have suggested that BE should be characterized by the sense of LOC rather than the quantity of food ingested (Hsu, Sullivan, & Benotti, 1997; Niego, Pratt, & Agras, 1997). Nonetheless, obese patients undergoing bariatric surgery and reporting LOC, even if that doesn’t include an objective large amount of food, seem to share with eating disordered patients elements of the psychopathological core of eating disorders (ED), presenting dysfunctional shape, weight and eating attitudes and high levels of psychosocial distress (de Zwaan, 2001; Hsu et al., 2002).
One point of interest concerns whether problems with BE, BED or LOC prior to surgery might be associated with less weight loss or more weight regain after surgery. As a corollary, researchers have also been interested in determining if such problems as BE, BED or LOC, may emerge or re-emerge after surgery, and then compromise weight outcomes. The first question has been addressed in a number of studies and in the majority of these studies BE, BED and LOC pre-surgery do not appear to be consistently predictive of weight outcomes (see, for example, de Zwaan et al., 2010; Hsu, Betancourt, & Sullivan, 1996; Kalarchian, Wilson, Brolin, & Bradley, 1999; and Powers, Perez, Boyd, & Rosemurgy, 1999), although some studies have reported worst outcomes for pre-surgery BE patients (Green et al., 2004; Sallet et al., 2007; Scholtz et al., 2007), and other studies reported that weight loss was greatest for those with pre-surgery BE (Boan, Kolotkin, Westman, McMahon, & Grant, 2004; Latner, Wetzler, Goodman, & Glinski, 2004; Malone & Alger-Mayer, 2004). However, the literature addressing the second question concerning post-surgery eating behaviour has yielded more consistent results. In this paper, we will restrict our focus to studies addressing BE/BED/LOC eating seen post-bariatric surgery, and review the literature on the prevalence and impact of post-surgery BE/BED/LOC eating on weight loss and weight regain. This literature was previously reviewed by Niego, Kofman, Weiss, and Geliebter in 2007.
A search by Pubmed, PsycInfo and Scopus was conducted using the terms Roux-en-Y anastamosis and obesity, bariatric surgery, obesity surgery, gastric bypass, gastroplasty, jejunoileal bypass and lipectomy. We combined these terms with BED, BE, bulimia nervosa and LOC. Dates searched began in April 2013 and ranged back to 1954 for PsycInfo, 1966 for Pubmed and no limits were used for Scopus. The study selection is shown in Figure 1.
The data extracted from the available literature are summarized in Table 1 and will only be briefly reviewed here. Studies dating back to 1992 have addressed this issue. Sample sizes have varied widely, from a minimum of 16 subjects to a maximum of 497 in the report by Kofman, Lent, and Swencionis (2010). A number of the samples have focused on patients who have undergone Roux-en-Y gastric bypass with a minority focusing on gastric bands and several early reports focusing on vertical banded gastroplasty. In one of the reports, different types of procedures were utilized (de Man Lapidoth, Ghaderi, & Norring, 2011). The durations of follow-up have also varied widely and frequently have included large ranges, with the longest follow-up study having been reported at 14 years by Mitchell et al. (2001). Various assessment strategies have been employed some using self-report and/or structured instruments such as the Eating Disorder Examination and the Structured Clinical Interview for the DSM. Of critical importance, the definitions of BE, BED (because this construct evolved over the time these publications appeared) and loss of control eating have varied across these studies, making comparisons difficult. This is likely reflected in the wide disparity in the prevalence of the rates of such behaviours in these reports, even among the studies with large sample sizes.
The prior review by Niego et al. published in 2007 included literature through 2006; therefore, approximately half of the studies cited in our current review would have been included. In this review, they found that the literature supported the notion that significant BE was related to poorer surgical outcome, and if such a behaviour was retained after surgery, the result was poorer weight loss outcomes.
In the only negative study reported to date, de Man Lapidoth et al. (2011) collected data from 4 surgery centres in Sweden. Various bariatric procedures were employed. One hundred two provided follow-up data, which revealed that 29 (28.4%) reported post-operative BE. This group did not differ from those not reporting this symptom on a three-year BMI outcome. BE after weight loss surgery was associated with psychopathology and lower self-reported health-related quality of life. The authors stressed that the occurrence of such problems suggests a need for surgical teams to be observant about the occurrence of these behaviours.
A review of these data show several important points: first, the disappearance of pre-surgery BE with surgery in a significant percentage of patients has been frequently described. About one-half of the studies evaluated pre-operative eating-disordered behaviour, with two studies examining what was labelled pre-operative LOC rather than BE, although LOC would have been assessed in other studies given that subjects satisfied criteria for BE. LOC eating episodes that were not BE episodes might have been underreported in the studies reporting only BE or BED. Second, all but one study suggest that the presence of BE/BED/LOC eating post-operatively is associated with less weight loss or more weight regain. Additionally, we see that of the studies that did evaluate post-operative BE/BED/LOC eating, rates varied widely from 0% in Latner et al. (2004) to 46% in the Kalarchian et al. paper (2002). Possible contributors to this wide variance include the different types of surgery employed (e.g. LAGB vs Roux-en-Y gastric bypass vs vertical banded gastroplasty), and the type of assessment (e.g. self-report measures vs face-to-face interview; QEWP-R vs EDE-Q), as well as length of time to follow-up. Additionally, in obesity, BE episodes are known to be difficult to assess because they are not as distinctive as such episodes in bulimia nervosa, where they are often terminated with some type of compensatory behaviour such as vomiting (de Zwaan et al., 2004; Marcus, Smith, Santelli, & Kaye, 1992). There are other potential problems with the LOC construct that require further clarification. For example, there are concerns that this label may be applied retrospectively by patients who assess that an eating episode was problematic. There also may be gradations of LOC, as opposed to viewing this characteristic as a dichotomous variable (Mitchell et al., 2012).
A possible interactive effect between of LOC and time has also been highlighted. Hsu and colleagues (1997) found that the initial improvement in pre-surgery ED status erodes after 2 years, significantly predicting poorer outcomes, and Sallet et al. (2007) noted that at two years’ post-surgery, the association between pre-surgery BE status and weight outcomes becomes more clearly manifested. These findings suggest that data collected with less than two years of follow-up may be misread. However, other work suggests value in having LOC assessed fairly early in follow-up. The prospective value of LOC eating was described by White and colleagues (2010). These authors found that LOC at 6 months significantly predicted weight loss at latter assessment times, and LOC at 12 months significantly predicted weight loss at 24 months.
Another factor that may account for the contradictory data is that different studies use different weight outcomes (BMI, BMI loss, % excess weight loss, weight regain), and that only a small group of studies assessed concomitant depressive symptoms and other potentially maladaptive eating behaviours such as grazing, nibbling, sweet eating or night eating syndrome. Although BE may be physically impossible for many patients due to surgical restriction of gastric volume, a different form of compulsive eating may emerge after surgery, and new maladaptive eating behaviours, like grazing (Saunders, 2001), rumination (chewing food, then spitting it out) as well as LOC (White, Kalarchian, Masheb, Marcus, & Grilo, 2010). These might appear to serve the same function as the previous BE (Saunders, 2004), resulting in excessive caloric intake and ultimately weight regain (van Hout et al., 2005).
The importance of considering the presence of LOC eating postoperatively seems to be clear in order to better understand weight loss and weight regain after bariatric surgery. However, important questions on different dimensions of this behaviour remain unanswered. Regarding the (re) emergence of BE/BED/LOC: why do some people experience BE/BED/LOC eating at times (re)emerging of this after surgery, whereas others do not? Is there a critical follow-up time for the emergence of these problems post-operatively? What are the predictive factors for the emergence or reemergence of these problems? Can clinicians distinguish, before surgery patients that will maintain their BE/BED/LOC status after surgery from those who will remit? Regarding the association with other dysfunctional eating behaviours or eating psychopathology: does pre-operative BE/BED/LOC eating develop into other eating behaviours (e.g. grazing) that also have an important impact on weight outcomes after surgery? Do such problems overlap with other maladaptive eating behaviours resulting in greater impact on weight loss/regain? Regarding treatment and health services: can clinicians prevent or treat the (re)emergence of BE/BED/LOC eating post-operatively? To our knowledge, no systematic work has been reported in this area.
In conclusion, the variability in weight outcomes after bariatric surgery is considerable, particularly at long-term follow-up, and dysfunctional eating appears to be associated with weight regain. However, the factors contributing to successful outcomes after surgery are complex, multiple and probably time-related, and the importance of single-predictive variables should not be over emphasized.
Because eating behaviours post-operatively are associated with weight loss and/or regain, attention should be focused on the follow-up period. Considering the prospective value of BE/BED/LOC eating, a preventive attitude based or early detection appear to be sensitive strategies to help prevent poor outcomes, and there is a clear need for studies focusing on prevention and treatment.
In summary, our findings agree with conclusions offered by Niego, Pratt and Agras in an earlier review (2007) suggesting that the emergence or re-emergence of BE/BED/LOC after bariatric surgery results in less weight loss and/or more weight regain. The available literature, with the exception of the study by de Man Lapidoth et al. (2011), support such an association.
Supported by RO1 DK 84979 (JEM) from the National Institute of Diabetes, Digestive and Kidney Diseases in the National Institutes of Health.